[AGENT++] Using multiple snmp++ instances on the same machine and timeouts

[Frank Fock]

Hi Hendrik,

Have you checked the size of the UDP buffer on your system.
If this is the bottle neck, either increasing it or using more retries
could solve the issue.

Best regards,
Frank

Hendrik-Willem van Zutphen wrote:

>Hi all,
>I'm using the SNMP++ library to monitor multiple devices (and fill a cache
>for use in a displaying application).
>After some performance tests the decision was made to use multiple instances
>of the SNMP class to be able to handle as much update requests as possible.
>In the current situation we have 1 SNMP instance handling trap and trap
>related requests and 17 "worker" SNMP instances for the other data. All the
>requests we are doing are multi VB get and multi VB bulk gets.
>
>When running the application I receive a lot of timeouts on requests done
>(which will be requested again eventually). The machine is using at most 10%
>CPU so nothing special there, is there an extra item needed when
>constructing the class to be able to run multiple instances, is there
>something else that I'm missing or is just the other side at fault (but
>timeouts of 8 seconds are rather long for a machine that is doing nothing
>but answering SNMP requests).
>
>Thanks in advance,
>
>Hendrik-Willem van Zutphen
>
>____________
>Artificial Intelligence usually beats real stupidity
>
>Hendrik-Willem van Zutphen
>hzut at oce.nl
>
>
>_______________________________________________
>AGENTPP mailing list
>AGENTPP at agentpp.org
>http://agentpp.org/mailman/listinfo/agentpp
>
>  
>